In the medical arena the use of inactive compounds such as prodrugs which are activated in a specific site in the human or animal body is well known. Also targeted delivery of inactives such as prodrugs has been studied extensively. Much effort has been devoted to drug delivery systems that effect drug release selectivity at a target site and/or at a desired moment in time. One way is to selectively activate a (systemic) prodrug specifically by local and specific enzymatic activity. However, in many cases a target site of interest lacks a suitable overexpressed enzyme. An alternative is to transport an enzyme to target tissue via a technique called antibody-directed enzyme prodrug therapy (ADEPT). In this approach an enzyme is targeted to a tumor site by conjugation to an antibody that binds a tumor-associated antigen. After systemic administration of the conjugate, its localization at the target and clearance of unbound conjugate, a designed prodrug is administered systemically and locally activated. This method requires the catalysis of a reaction that must not be accomplished by an endogenous enzyme. Enzymes of non-mammalian origin that meet these needs are likely to be highly immunogenic, a fact that makes repeated administration impossible. Alternatively, prodrugs can be targeted to a disease site followed by disease-specific or -non-specific endogenous activation processes (eg pH, enzymes, thiol-containing compounds).
Targeted anticancer therapeutics are designed to reduce nonspecific toxicities and increase efficacy relative to conventional cancer chemotherapy. This approach is embodied by the powerful targeting ability of monoclonal antibodies (mAbs) to specifically deliver highly potent, conjugated small molecule therapeutics to a cancer cell. In an attempt to address the issue of toxicity, chemotherapeutic agents (drugs) have been coupled to targeting molecules such as antibodies or protein receptor ligands that bind with a high degree of specificity to tumor cell to form compounds referred to as antibody-drug conjugates (ADC) or immunoconjugates. Immunoconjugates in theory should be less toxic because they direct the cytotoxic drug to tumors that express the particular cell surface antigen or receptor. This strategy has met limited success in part because cytotoxic drugs tend to be inactive or less active when conjugated to large antibodies, or protein receptor ligands. Promising advancements with immunoconjugates has seen cytotoxic drugs linked to antibodies through a linker that is cleaved at the tumor site or inside tumor cells (Senter et al, Current Opinion in Chemical Biology 2010, 14:529-537). Ideally, the mAb will specifically bind to an antigen with substantial expression on tumor cells but limited expression on normal tissues. Specificity allows the utilization of drugs that otherwise would be too toxic for clinical application. Most of the recent work in this field has centered on the use of highly potent cytotoxic agents. This requires the development of linker technologies that provide conditional stability, so that drug release occurs after tumor binding, rather than in circulation.
As a conjugate the drug is inactive but upon target localization the drug is released by eg pH or an enzyme, which could be target specific but may also be more generic. The drug release may be achieved by an extracellular mechanism such as low pH in tumor tissue, hypoxia, certain enzymes, but in general more selective drug release can be achieved through intracellular, mostly lysosomal, release mechanisms (e.g. glutathione, proteases, catabolism) requiring the antibody conjugate to be first internalized. Specific intracellular release mechanisms (eg glutathione, cathepsin) usually result in the parent drug, which depending on its properties, can escape the cell and attack neighboring cells. This is viewed as an important mechanism of action for a range of antibody-drug conjugates, especially in tumors with heterogeneous receptor expression, or with poor mAb penetration. Examples of cleavable linkers are: hydrazones (acid labile), peptide linkers (cathepsin B cleavable), hindered disulfide moieties (thiol cleavable). Also non-cleavable linkers can be used in mAb-drug conjugates. These constructs release their drug upon catabolism, presumably resulting in a drug molecule still attached to one amino acid. Only a subset of drugs will regain their activity as such a conjugate. Also, these aminoacid-linked drugs cannot escape the cells. Nevertheless, as the linker is stable, these constructs are generally regarded as the safest and depending on the drug and target, can be very effective.
The current antibody-drug conjugate release strategies have their limitations. The extracellular drug release mechanisms are usually too unspecific (as with pH sensitive linkers) resulting in toxicity. Intracellular release depends on efficient (e.g receptor-mediated internalization) of the mAb-drug, while several cancers lack cancer-specific and efficiently internalizing targets that are present in sufficiently high copy numbers. Intracellular release may further depend on the presence of an activating enzyme (proteases) or molecules (thiols such as glutathione) in sufficiently high amount. Following intracellular release, the drug may, in certain cases, escape from the cell to target neighbouring cells. This effect is deemed advantageous in heterogeneous tumors where not every cell expresses sufficiently high amounts of target receptor. It is of further importance in tumors that are difficult to penetrate due e.g. to elevated interstitial pressure, which impedes convectional flow. This is especially a problem for large constructs like mAb (conjugates). This mechanism is also essential in cases where a binding site barrier occurs. Once a targeted agent leaves the vasculature and binds to a receptor, its movement within the tumor will be restricted. The likelihood of a mAb conjugate being restricted in the perivascular space scales with its affinity for its target. The penetration can be improved by increasing the mAb dose, however, this approach is limited by dose limiting toxicity in e.g. the liver. Further, antigens that are shed from dying cells can be present in the tumor interstitial space where they can prevent mAb-conjugates of binding their target cell. Also, many targets are hampered by ineffective internalization, and different drugs cannot be linked to a mAb in the same way. Further, it has been proven cumbersome to design linkers to be selectively cleavable by endogenous elements in the target while stable to endogenous elements en route to the target (especially the case for slow clearing full mAbs). As a result, the optimal drug, linker, mAb, and target combination needs to be selected and optimized on a case by case basis.
Another application area that could benefit from an effective prodrug approach is the field of T-cell engaging antibody constructs (e.g., bi- or trispecific antibody fragments), which act on cancer by engaging the immune system. It has long been considered that bringing activated T-cells into direct contact with cancer cells offers a potent way of killing them (Thompson et al., Biochemical and Biophysical Research Communications 366 (2008) 526-531). Of the many bispecific antibodies that have been created to do this, the majority are composed of two antibody binding sites, one site targets the tumor and the other targets a T-cell (Thakur et al. Current Opinion in Molecular Therapeutics 2010, 12(3), 340-349). However, with bispecific antibodies containing an active T-cell binding site, peripheral T-cell binding will occur. This not only prevents the conjugate from getting to the tumor but can also lead to cytokine storms and T-cell depletion. Photo-activatable anti-T-cell antibodies, in which the anti-T-cell activity is only restored when and where it is required (i.e. after tumor localization via the tumor binding arm), following irradiation with UV light, has been used to overcome these problems. Anti-human CD3 (i-cell targeting) antibodies could be reversibly inhibited with a photocleavable 1-(2-nitrophenyl)ethanol (NPE) coating (Thompson et al., Biochemical and Biophysical Research Communications 366 (2008) 526-531). However, light based activation is limited to regions in the body where light can penetrate, and is not easily amendable to treating systemic disease such as metastatic cancer. Strongly related constructs that could benefit from a prodrug approach are trispecific T-cell engaging antibody constructs with for example a CD3- and a CD28 T-cell engaging moiety in addition to a cancer targeting agent. Such constructs are too toxic to use as such and either the CD3 or the CD28 or both binding domains need to be masked.
It is desirable to be able to activate targeted drugs selectively and predictably at the target site without being dependent on homogenous penetration and targeting, and on endogenous parameters which may vary en route to and within the target, and from indication to indication and from patient to patient.
In order to avoid the drawbacks of current prodrug activation, it has been proposed in Bioconjugate Chem 2008, 19, 714-718, to make use of an abiotic, bio-orthogonal chemical reaction, viz. the Staudinger reaction, to provoke activation of the prodrug. Briefly, in the introduced concept, the Prodrug is a conjugate of a Drug and a Trigger, and this Drug-Trigger conjugate is not activated endogeneously by e.g. an enzyme or a specific pH, but by a controlled administration of the Activator, i.e. a species that reacts with the Trigger moiety in the Prodrug, to induce release of the Drug from the Trigger (or vice versa, release of the Trigger from the Drug, however one may view this release process). The presented Staudinger approach for this concept, however, has turned out not to work well, and its area of applicability is limited in view of the specific nature of the release mechanism imposed by the Staudinger reaction. Other drawbacks for use of Staudinger reactions are their limited reaction rates, and the oxidative instability of the phosphine components of these reactions. Therefore, it is desired to provide reactants for an abiotic, bio-orthogonal reaction that are stable in physiological conditions, that are more reactive towards each other, and that are capable of inducing release of a bound drug by means of a variety of mechanisms, thus offering a greatly versatile activated drug release method.
The use of a biocompatible chemical reaction that does not rely on endogenous activation mechanisms (eg pH, enzymes) for selective Prodrug activation would represent a powerful new tool in cancer therapy. Selective activation of Prodrugs when and where required allows control over many processes within the body, including cancer. Therapies, such as anti-tumor antibody therapy, may thus be made more specific, providing an increased therapeutic contrast between normal cells and tumour to reduce unwanted side effects. In the context of T-cell engaging anticancer antibodies, the present invention allows the systemic administration and tumor targeting of an inactive antibody construct (i.e. this is then the Prodrug), diminishing off-target toxicity. Upon sufficient tumor uptake and clearance from non target areas, the tumor-bound antibody is activated by administration of the Activator, which reacts with the Trigger or Triggers on the antibody or particular antibody domain, resulting in removal of the Trigger and restoration of the T-cell binding function. This results in T-cell activation and anticancer action (i.e. this is then the Drug release).